NVIDIA GeForce GTX 1070 Ti vs NVIDIA GeForce RTX 2060 Comparison

The NVIDIA GeForce RTX 2060 and the NVIDIA GeForce GTX 1070 Ti are evenly matched in the benchmark data, with each card winning five of the ten head-to-head tests. However, the nature of those wins reveals a clear split: the RTX 2060 dominates in modern compute and API-specific workloads, while the GTX 1070 Ti holds advantages in raw rasterization and legacy DirectX performance. The average benchmark scores reflect this, with the RTX 2060 averaging 15290 and the GTX 1070 Ti averaging 14277, a difference of roughly 7% in favor of the newer card. The RTX 2060 also sits at the 58th percentile of all GPUs, two points higher than the GTX 1070 Ti's 56th percentile.

Head-to-Head Benchmarks

The most decisive victory belongs to the RTX 2060 in the Geekbench OpenCL test, where it scores 65014 against the GTX 1070 Ti's 49219. That is a 32.1% advantage, the largest delta in either direction across all ten benchmarks. This is a massive gap that points directly to the Turing architecture's strengths in general-purpose compute. The RTX 2060 also wins the Geekbench Vulkan test, scoring 65846 versus 61006, a 7.9% lead that shows its advantage extends into modern graphics APIs.

The GTX 1070 Ti strikes back hardest in the Passmark GPU Compute test, delivering a score of 7165 against the RTX 2060's 5488. That is a 23.4% margin, the second-largest delta in the entire comparison. This is a surprising result given the RTX 2060's OpenCL dominance, but the Pascal card clearly has specialized compute strengths that the Turing card does not match. The GTX 1070 Ti also wins the 3DMark Steel Nomad DX12 test with 1471 versus 1242, a 15.6% lead, which is significant for a modern DirectX 12 benchmark.

In the legacy DirectX 9 test, the GTX 1070 Ti scores 207 against 190 for the RTX 2060, an 8.2% win. The GTX 1070 Ti also takes the Passmark G2D test with 876 versus 745, a 15% lead, and the Passmark G3D test with 14674 versus 14111, a 3.8% margin. The RTX 2060 counters with narrower wins in the newer DirectX tests: 8.9% in DirectX 10 (98 vs 90), 2.8% in DirectX 11 (110 vs 107), and 3.9% in DirectX 12 (53 vs 51). The pattern is consistent: the GTX 1070 Ti excels in older and rasterization-focused workloads, while the RTX 2060 pulls ahead in compute-heavy and API-agnostic tests.

FAQ

Q: Which card has a higher average benchmark score?

A: The NVIDIA GeForce RTX 2060 has an average benchmark score of 15290, compared to 14277 for the NVIDIA GeForce GTX 1070 Ti.

Q: In which benchmark does the GTX 1070 Ti have its largest advantage?

A: The GTX 1070 Ti wins the Passmark GPU Compute test by 23.4%, scoring 7165 against the RTX 2060's 5488.

Q: What is the biggest win for the RTX 2060?

A: The RTX 2060 wins the Geekbench OpenCL test by 32.1%, with a score of 65014 versus 49219 for the GTX 1070 Ti.

Q: Which card performs better in the 3DMark Steel Nomad DX12 test?

A: The GTX 1070 Ti leads in 3DMark Steel Nomad DX12, scoring 1471 compared to 1242 for the RTX 2060, a 15.6% difference.

Q: How do the cards compare in DirectX 11 performance?

A: The RTX 2060 scores 110 in Passmark DirectX 11, while the GTX 1070 Ti scores 107, giving the RTX 2060 a 2.8% edge.

Q: Does the GTX 1070 Ti win any modern API benchmarks?

A: Yes, the GTX 1070 Ti wins the 3DMark Steel Nomad DX12 test with 1471 versus 1242, a 15.6% lead over the RTX 2060.

Architecture Differences

The two cards are built on fundamentally different architectures. The RTX 2060 uses the TU106 chip based on the Turing architecture, fabricated on a 12 nm process at TSMC. The GTX 1070 Ti uses the GP104 chip based on the older Pascal architecture, fabricated on a 16 nm process, also at TSMC. The Turing chip packs 10,800 million transistors into a 445 mm² die, while the Pascal chip contains 7,200 million transistors on a 314 mm² die. This results in a transistor density of 24.3M per mm² for the RTX 2060, versus 22.9M per mm² for the GTX 1070 Ti.

The most significant architectural difference is the inclusion of dedicated ray tracing and tensor cores on the RTX 2060. It features 30 RT cores and 240 tensor cores, neither of which exist on the GTX 1070 Ti. This is the defining feature of the GeForce 20-series, enabling hardware-accelerated ray tracing and AI-driven features that the Pascal card simply cannot execute. The RTX 2060 also has a different FP16 implementation, delivering 12.90 TFLOPS with a 2:1 ratio, while the GTX 1070 Ti manages only 127.9 GFLOPS with a 1:64 ratio. This makes the RTX 2060 vastly more capable in half-precision compute workloads.

The shading unit configuration also differs. The GTX 1070 Ti has 2432 shading units, 152 texture mapping units, and 64 ROPs, while the RTX 2060 has 1920 shading units, 120 TMUs, and 48 ROPs. Despite having fewer shading units, the RTX 2060's Turing architecture uses a different scheduling and execution model that improves efficiency per unit. The pixel rate tells the story: the GTX 1070 Ti achieves 107.7 GPixel/s versus 80.64 GPixel/s for the RTX 2060, while the texture rate is 255.8 GTexel/s versus 201.6 GTexel/s. The Pascal card has a clear raw throughput advantage in these traditional metrics.

Specification Differences

The memory subsystems are notably different. The RTX 2060 uses 6 GB of GDDR6 memory on a 192-bit bus, delivering 336.0 GB/s of bandwidth. The GTX 1070 Ti uses 8 GB of GDDR5 memory on a wider 256-bit bus, but achieves only 256.3 GB/s of bandwidth. The RTX 2060 has the bandwidth advantage despite the narrower bus, thanks to the faster GDDR6 memory running at 14 Gbps effective, versus 8 Gbps effective for the GTX 1070 Ti. The GTX 1070 Ti offers 2 GB more memory capacity, which could matter in texture-heavy scenarios.

Clock speeds are close but favor the Pascal card. The GTX 1070 Ti has a base clock of 1607 MHz and a boost clock of 1683 MHz, while the RTX 2060 has a base clock of 1365 MHz and a boost clock of 1680 MHz. The GTX 1070 Ti runs at a higher base frequency, though the boost clocks are nearly identical. The power draw differs, with the GTX 1070 Ti rated at 180 W TDP versus 160 W for the RTX 2060, despite the Pascal card using an older process node.

Physical dimensions also set them apart. The RTX 2060 is shorter at 229 mm (9 inches) versus 267 mm (10.5 inches) for the GTX 1070 Ti, and thinner at 35 mm (1.4 inches) versus 40 mm (1.6 inches). Both are dual-slot cards with a single 8-pin power connector and a 450 W suggested PSU. The display outputs differ: the RTX 2060 includes a USB Type-C port alongside its DVI, HDMI 2.0, and two DisplayPort 1.4a outputs, while the GTX 1070 Ti has three DisplayPort 1.4a outputs instead of the USB Type-C. The RTX 2060 supports DirectX 12 Ultimate (12_2), while the GTX 1070 Ti is limited to DirectX 12 (12_1).

Where Each One Wins

The GTX 1070 Ti is the stronger choice for traditional rasterization workloads. It wins the 3DMark Steel Nomad DX12 test by 15.6%, the Passmark G3D test by 3.8%, and the Passmark DirectX 9 test by 8.2%. Its higher pixel rate of 107.7 GPixel/s and texture rate of 255.8 GTexel/s give it a raw fill-rate advantage that shows up in these benchmarks. For users running older DirectX 9 titles or games that rely heavily on classic rasterization, the GTX 1070 Ti's 8 GB of GDDR5 memory and 256-bit bus provide a solid foundation. The Passmark G2D result, where it leads by 15%, also suggests better 2D performance for desktop and productivity scenarios.

The RTX 2060 is the winner in compute-heavy and modern API workloads. Its 32.1% lead in Geekbench OpenCL and 7.9% lead in Geekbench Vulkan demonstrate a clear edge in general-purpose GPU compute and the Vulkan API. The 240 tensor cores enable AI-accelerated features that the GTX 1070 Ti cannot match, and the 30 RT cores provide hardware ray tracing support. In the Passmark DirectX 10, 11, and 12 tests, the RTX 2060 wins all three with margins of 8.9%, 2.8%, and 3.9% respectively, showing consistent superiority across DirectX versions. Its 336.0 GB/s of memory bandwidth, despite a smaller 6 GB pool, helps in bandwidth-sensitive compute tasks.

The Verdict

The data supports a clear split decision. The RTX 2060 is the better all-around card for modern workloads, evidenced by its higher average benchmark score of 15290 and its wins in OpenCL, Vulkan, and all three Passmark DirectX tests (10, 11, and 12). Its 58th percentile ranking versus 56th for the GTX 1070 Ti confirms its slight overall edge. Buyers who prioritize compute performance, modern API support, and the architectural features of ray tracing and tensor cores should choose the RTX 2060, especially given its lower 160 W TDP and shorter physical footprint.

The GTX 1070 Ti is the pick for users focused on traditional gaming performance and legacy compatibility. It wins the 3DMark Steel Nomad DX12 test by 15.6%, the Passmark G3D test by 3.8%, and the Passmark GPU Compute test by a massive 23.4%. Its 8 GB of memory capacity is also larger, and its higher base clock of 1607 MHz gives it a head start in lighter loads. The 267 mm length is longer, but the card still fits standard dual-slot configurations. For those who do not need ray tracing or tensor core features and prioritize raw rasterization speed and compute in specific workloads, the GTX 1070 Ti remains a viable option despite its older Pascal architecture. The final call depends on whether the user values modern compute features or traditional rendering performance.